This is the 3rd update to my massive diary on the impact of gerrymandering nationwide, which strongly suggested that it cost Democrats the house. In case you missed it, the first update was to add Connecticut, and modestly improve Louisiana, Michigan, and New Jersey, while the 2nd was a drastically improved Illinois. This time though I felt like a stand alone diary was worthwhile because the issue of VRA interpretation was a systematic decision on my part that some might not have agreed with. Here I wanted to show that even if you don't force a second VRA district in Alabama and Louisiana like I believe ought to have been, you still get a second solidly Democratic district thanks to geographic racial polarization and the clustering of the few white Democrats in those states in cities. I'll eventually do South Carolina but haven't because it requires more time than the other two although it will very likely arrive at a similar conclusion. So let's take a look at the two maps and you can find the 2VRA district ones in the original diary.
Alabama
As you can see, it was quite easy to create one VRA black majority district that is confined entirely to the Black Belt region of Alabama and is anchored by Montgomery. The district only splits 3 counties and would be safe for an additional Democrat as an open seat in 2012. This allows me to take Birmingham and nearly every single one of its suburbs and place them in the same district. Obama won it by almost exactly his national margin and it would have been perfectly safe for incumbent Teri Sewell in presidential years, but she would have to work for it in a midterm. It's theoretically possible that the district could nominate a white liberal if open, but the primary electorate is going to be overwhelmingly Black and Alabama has runoffs. You could alternatively draw it so that the 6th takes in all of Jefferson county and a tiny bit of Shelby, but since it had to have some of the latter anyway I figured I'd try to drop all of the rural areas and take in as many suburbs as possible. None of the other districts is really competitive, with the least red being the 5th which is about the same partisanship as the state.
This was an improvement upon my previous one VRA district map which maintained the Black Belt to Birmingham configuration and allowed the Montgomery district to be a blue dog seat for Bobby Bright, but there's absolutely no reason why an independent commission would prefer that one over this. There were also minor improvements to the 5th district by restoring Decatur.
Louisiana
Starting with the premise that drawing a Baton Rouge to Shreveport district can't be forced, we go back to the drawing board. If you try to maintain the New Orleans based 2nd district as the VRA seat, you're required to shred apart both the greater Baton Rouge and New Orleans areas and use water contiguity across Lake Pontchartrain, both of which are quite unnecessary. The old 2nd district was contained entirely within the New Orleans area and it just so happens that you can fit almost the entirety of the metro area in one compact district. That allows the Baton Rouge based 6th to become a fairly compact VRA district without even setting foot in New Orleans. A bonus is that racial polarization is much higher here anyway compared to the New Orleans area. As such, the result is two still safe Democratic districts, but the New Orleans one now might elect a white liberal like Mitch Landrieu, the current very popular mayor although incumbent Cedric Richmond is probably still heavily favored. From a purely electoral standpoint, this map is actually slightly better than the 2VRA one because the 4th district in the north of the state is now theoretically winnable whereas before none of the others were. Finally, instead of giving then Rep. Landry the short end of the stick like the actual map did, his seat is revived as the Acadiana based 3rd and 4th district Rep. John Fleming keeps his seat. In exchange, Charles Boustany is thrown in at a disadvantage with 5th district Rep. Rodney Alexander while Reps Steve Scalise and Bill Cassidy would both run in the 1st unless Cassidy retired early for his 2014 senate run.
The only thing I really don't like about this district is the 5th and that's a result of Louisiana being a hell of a lot cleaner to draw with 7 districts rather than 6 as Lake Charles doesn't really belong with those Parishes along the Delta, but there's nowhere else for either to go. The only real alternative to this map is as I mentioned before a New Orleans to Baton Rouge VRA district that would basically be like the actual one, but there's no reason to significantly divide both metropolitan areas unless you're making a 2nd VRA district. Nearly all of the 2002-2010 2nd district is in this new one while the 1st drops the whiter parts of New Orleans for the whiter parts of Baton Rouge.
I should hopefully have a South Carolina map completed soon and my initial draft resulted in a Columbia district very similar in partisanship to the Birmingham Alabama one meaning that it would be slightly D+ which is pretty safe in the south but not for a hard core liberal. In addition to that, there would be an entirely rural/non-Charleston or Columbia black majority district, and a tossup/Lean R Charleston-Beaufort district where Tim Scott would have seen a close race if not lost and if Haley had still appointed him to the senate, Colbert Busch would have destroyed Mark Sanford. It'll take a little while to refine the map though so I didn't include it in this diary as I really don't want to run the numbers more than once.